基于神经网络的三端口DC-DC变换器解耦控制策略
|
摘要
针对隔离型三端口有源全桥变换器(TAB)存在的实用性问题,提出一种基于神经网络的解耦控制策略。传统的TAB解耦方法主要包括硬件解耦和软件解耦,硬件解耦通过改变电路拓扑消除耦合,工作模式相对固定,而软件解耦则相对灵活。但软件解耦控制需用查表法来获得与系统运行点密切相关的解耦矩阵,该方法在系统工况复杂时存在表格维数过大、实用性不高的问题。针对该问题,提出了基于神经网络的解耦矩阵在线计算方法,有效地降低了解耦算法的空间复杂度。其次,传统的TAB恒压端口采用端口直流内环加端口电压外环的控制方法,其中端口直流电流内环用于实现端口间的解耦控制。该方法中控制回路的设计依赖于负载动态特性,在负载动态特性未知时控制器无法设计。针对该问题,采用端口电压单环控制,并利用端口直流电流的低频分量来计算解耦矩阵,提高了解耦控制算法对不同负载的适应性。通过对负载变化的情况进行仿真,仿真结果表明,该方法消除了端口间的耦合关系,提高了系统的动态响应速度。
To solve the practical problem of isolated active three-port bridge(TAB) DC-DC converter, a decoupling control strategy based on neural network is proposed. The traditional TAB decoupling methods mainly include hardware decoupling and software decoupling. Hardware decoupling eliminates coupling by changing the circuit topology. So, the working mode is relatively fixed, while software decoupling is relatively flexible. However, the traditional TAB decoupling control method uses the table look-up method to obtain a decoupling matrix which is closely related to the system operating point. This method has the problem of too large table dimensions and low practicality when the system operating conditions are complex. To solve this problem, this paper proposes a decoupling matrix online calculation method based on neural network, which can effectively reduce the space complexity of the decoupling algorithm. Moreover, the conventional TAB constant voltage port adopts the control method of the port DC inner loop plus the port voltage outer loop, wherein the inner loop of the port DC current is used to realize the decoupling control between the ports. The design of the control loop in this method depends on the dynamic characteristics of the load. The controller cannot be designed when the dynamic characteristics of the load are unknown. To solve this problem, this paper uses port voltage single-loop control and uses the low-frequency components of the port DC current to calculate the decoupling matrix, which improves the adaptability of the decoupling control algorithm to different loads. Through the simulation of the load changes, the simulation results show that this method eliminates the coupling between ports and improves the system's dynamic response speed.
To solve the practical problem of isolated active three-port bridge(TAB) DC-DC converter, a decoupling control strategy based on neural network is proposed. The traditional TAB decoupling methods mainly include hardware decoupling and software decoupling. Hardware decoupling eliminates coupling by changing the circuit topology. So, the working mode is relatively fixed, while software decoupling is relatively flexible. However, the traditional TAB decoupling control method uses the table look-up method to obtain a decoupling matrix which is closely related to the system operating point. This method has the problem of too large table dimensions and low practicality when the system operating conditions are complex. To solve this problem, this paper proposes a decoupling matrix online calculation method based on neural network, which can effectively reduce the space complexity of the decoupling algorithm. Moreover, the conventional TAB constant voltage port adopts the control method of the port DC inner loop plus the port voltage outer loop, wherein the inner loop of the port DC current is used to realize the decoupling control between the ports. The design of the control loop in this method depends on the dynamic characteristics of the load. The controller cannot be designed when the dynamic characteristics of the load are unknown. To solve this problem, this paper uses port voltage single-loop control and uses the low-frequency components of the port DC current to calculate the decoupling matrix, which improves the adaptability of the decoupling control algorithm to different loads. Through the simulation of the load changes, the simulation results show that this method eliminates the coupling between ports and improves the system's dynamic response speed.
引文
[1] 李子欣,高范强,赵聪,等(Li Zixin, Gao Fanqiang, Zhao Cong, et al.). 电力电子变压器技术研究综述(Research review of power electronic transformer technologies)[J].中国电机工程学报(Proceedings of the CSEE),2018,38(5):1274-1289.
[2] 孙孝峰,薛利,孟宇飞, 等(Sun Xiaofeng, Xue Li, Meng Yufei,et al.). 双串联谐振双向三端口DC/DC变换器选频解耦控制研究(Power decoupling control scheme with selected frequency for a dual series resonant bidirectional three-port DC/DC converter)[J]. 太阳能学报(Acta Renewable Solaris Sinica),2015,36(6):1459-1467.
[3] Su G J, Tang L. A reduced-part, triple-voltage DC-DC converter for EV/HEV power management[J]. IEEE Transactions on Power Electronics, 2009, 24(10):2406-2410.
[4] Chen Y M, Liu Y C, Lin S H. Double-input PWM DC/DC converter for high-low-voltage sources[J]. IEEE Transactions on Industrial Electronics, 2006, 53(5):1538-1545.
[5] 陈润若, 吴红飞, 邢岩(Chen Runruo, Wu Hongfei, Xing Yan). 一种适用于宽输入电压范围的三端口变换器(A novel three-port converter for wide-input-voltage-range application)[J]. 中国电机工程学报(Proceedings of the CSEE),2012,32(27):119-125.
[6] Zhao C, Round S D, Kolar J W. An isolated three-port bidirectional DC-DC converter with decoupled power flow management [J]. IEEE Transactions on Power Electronics, 2008, 23(5):2443-2453.
[7] Krishnaswami H, Mohan N. Three-port series-resonant DC-DC converter to interface renewable energy sources with bidirectional load and energy storage ports[J]. IEEE Transactions on Power Electronics, 2009, 24(10):2289-2297.
[8] 杨旭, 王卫, 王盼宝,等(Yang Xu,Wang Wei, Wang Panbao, et al.). 基于串联谐振网络的三端口DC/DC变换器解耦方法(Decoupling method of Series resonance based three-port DC/DC converter)[J]. 电网技术(Power System Technology),2017,41(2):478-485.
[9] Xu D, Zhao C, Fan H. A PWM plus phase-shift control bidirectional DC-DC converter[J]. Proceedings of the CSEE, 2003, 19(3):666-675.
[10] Ganz A G. A simple, exact equivalent circuit for the three-winding transformer[J]. IRE Transactions on Component Parts, 1962, 9(4):212-213.
[11] Cun Y L, Boser B, Denker J S, et al. Handwritten digit recognition with a back-propagation network[J]. Advances in Neural Information Processing Systems, 1990, 2(2):396-404.
[12] Kivinen J, Smola A J, Williamson R C. Online learning with kernels[J]. IEEE Transactions on Signal Processing, 2004, 52(8):2165-2176.
[2] 孙孝峰,薛利,孟宇飞, 等(Sun Xiaofeng, Xue Li, Meng Yufei,et al.). 双串联谐振双向三端口DC/DC变换器选频解耦控制研究(Power decoupling control scheme with selected frequency for a dual series resonant bidirectional three-port DC/DC converter)[J]. 太阳能学报(Acta Renewable Solaris Sinica),2015,36(6):1459-1467.
[3] Su G J, Tang L. A reduced-part, triple-voltage DC-DC converter for EV/HEV power management[J]. IEEE Transactions on Power Electronics, 2009, 24(10):2406-2410.
[4] Chen Y M, Liu Y C, Lin S H. Double-input PWM DC/DC converter for high-low-voltage sources[J]. IEEE Transactions on Industrial Electronics, 2006, 53(5):1538-1545.
[5] 陈润若, 吴红飞, 邢岩(Chen Runruo, Wu Hongfei, Xing Yan). 一种适用于宽输入电压范围的三端口变换器(A novel three-port converter for wide-input-voltage-range application)[J]. 中国电机工程学报(Proceedings of the CSEE),2012,32(27):119-125.
[6] Zhao C, Round S D, Kolar J W. An isolated three-port bidirectional DC-DC converter with decoupled power flow management [J]. IEEE Transactions on Power Electronics, 2008, 23(5):2443-2453.
[7] Krishnaswami H, Mohan N. Three-port series-resonant DC-DC converter to interface renewable energy sources with bidirectional load and energy storage ports[J]. IEEE Transactions on Power Electronics, 2009, 24(10):2289-2297.
[8] 杨旭, 王卫, 王盼宝,等(Yang Xu,Wang Wei, Wang Panbao, et al.). 基于串联谐振网络的三端口DC/DC变换器解耦方法(Decoupling method of Series resonance based three-port DC/DC converter)[J]. 电网技术(Power System Technology),2017,41(2):478-485.
[9] Xu D, Zhao C, Fan H. A PWM plus phase-shift control bidirectional DC-DC converter[J]. Proceedings of the CSEE, 2003, 19(3):666-675.
[10] Ganz A G. A simple, exact equivalent circuit for the three-winding transformer[J]. IRE Transactions on Component Parts, 1962, 9(4):212-213.
[11] Cun Y L, Boser B, Denker J S, et al. Handwritten digit recognition with a back-propagation network[J]. Advances in Neural Information Processing Systems, 1990, 2(2):396-404.
[12] Kivinen J, Smola A J, Williamson R C. Online learning with kernels[J]. IEEE Transactions on Signal Processing, 2004, 52(8):2165-2176.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |